CN113160252B - Hierarchical segmentation method for cultural pattern image - Google Patents
Hierarchical segmentation method for cultural pattern image Download PDFInfo
- Publication number
- CN113160252B CN113160252B CN202110563186.1A CN202110563186A CN113160252B CN 113160252 B CN113160252 B CN 113160252B CN 202110563186 A CN202110563186 A CN 202110563186A CN 113160252 B CN113160252 B CN 113160252B
- Authority
- CN
- China
- Prior art keywords
- pixel
- representing
- calculating
- formula
- adjacent
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000011218 segmentation Effects 0.000 title claims abstract description 28
- 239000003086 colorant Substances 0.000 claims abstract description 5
- 238000004364 calculation method Methods 0.000 claims description 18
- 230000005284 excitation Effects 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 230000004075 alteration Effects 0.000 claims 1
- 230000006835 compression Effects 0.000 abstract description 6
- 238000007906 compression Methods 0.000 abstract description 6
- 238000011160 research Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 230000001364 causal effect Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000009940 knitting Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a hierarchical segmentation method for a cultural pattern image, which is proposed based on compression and merging theory, wherein the image is segmented into a plurality of initial areas by calculating color similarity characteristics among pixels, and the areas are merged by calculating colors, textures, area sizes and interweaving degrees among the areas to obtain a hierarchical segmentation result. Compared with the existing segmentation method, the method has the advantages that the segmentation speed is high, more image details can be reserved as a result, and simultaneously, cultural patterns with different detail degrees can be obtained.
Description
Technical Field
The invention belongs to the field of image processing and computer vision, and particularly relates to a hierarchical segmentation method for cultural pattern images.
Background
As the Chinese ethnic excellent traditional culture is becoming popular, the research of traditional culture is gradually paid attention to, and the interpretation of ethnic historical development and custom culture becomes the beginning direction of research. The method aims to solve the problem that cultural patterns are obtained by storing cultural relics, frescoes and the like containing the cultural patterns as image data through photographing, scanning and other modes and dividing the images. Compared with a natural image, the cultural pattern image has the characteristics of complex texture, unsmooth color area and many noise points, and when the conventional image segmentation method is used for segmenting the cultural pattern image, the pattern with complete edges cannot be extracted, but the method is continuously optimized aiming at the characteristics of the cultural pattern image, and the segmentation result with complete edges can be obtained.
Disclosure of Invention
The invention aims to solve the problems in the prior art, and provides a hierarchical segmentation method for a cultural pattern image, which is used for segmenting one image into a plurality of results with hierarchical relationship, wherein each layer of segmentation result accords with causal and inclusion principles, and the automatic segmentation of the cultural pattern image is completed, so that all elements in the image are obtained.
The invention provides a hierarchical segmentation method for cultural pattern images, which is characterized by comprising the following steps of:
step 1, converting a cultural pattern image into a Lab color space, and compressing pixels with similar color characteristics together, so that the image is divided into a plurality of areas;
step 2, calculating likelihood ratios between adjacent areas based on color features, and combining the adjacent areas with the likelihood ratios larger than a preset threshold value to obtain an over-segmentation result, wherein a likelihood ratio calculation formula is as follows:
wherein i and j represent the sequence numbers of the regions, R i and Rj Representing the area of the neighborhood,
representation area R i Is used for the color value of the color filter,
representation area R j Mean color value of>
Mean color value representing boundary between adjacent regions, S representing the color represented by adjacent region R i And R is R j Covariance matrix of pixel color values of the composed region;
step 3, calculating the difference of adjacent areas
And 3.1, calculating the average difference of the colors between the adjacent areas and the average difference of the colors at the boundaries of the adjacent areas, wherein the calculation formulas are respectively as follows:
in the formula,Ri and Rj Representing the area of the neighborhood,
representation area R i Pixel average color value of>
Representation area R j Pixel average color value of B ij Representing the boundary between adjacent regions, p and q representing the pixels of the regions on either side of the boundary, W p Sliding window representing center pixel p, B C (R i ,R j ) The number of (p, q) pairs representing the boundary, < ->
Respectively representing the brightness of the pixel p in Lab color space, the component from green to red, the component from blue to yellow,/respectively>
Respectively representing the brightness of the pixel q in the Lab color space, the component from green to red, and the component from blue to yellow;
and 3.2, calculating texture differences between adjacent areas, wherein the calculation formula is as follows:
D T (R i ,R j )＝D AB (R i ,R j )*D W (R i ,R j )
in the formula,Ri and Rj Representing adjacent regions, D AB (R i ,R j ) Representing the chromaticity difference of adjacent regions, D W (R i ,R j ) Representing texture feature differences for adjacent regions;
3.3, calculating the scale difference between adjacent areas, wherein the calculation formula is as follows:
wherein ,
representation area R i The number of pixels in +.>
Representation area R j The number of pixels in;
and 3.4, calculating an interleaving value between adjacent areas, wherein the calculating formula is as follows:
in the formula,Np Representing the most pixel color value, N, within a sliding window centered on pixel p q Representing the most pixel color value, C, within a sliding window centered on pixel q j Representation area R j Color value of the pixel with the largest inner part, C i Representation area R i The pixel color value with the largest inner, when a=b, δ (a, b) =1; when a+.b, δ (a, b) =0;
and 3.5, calculating the comprehensive difference between adjacent areas, wherein the calculation formula is as follows:
wherein α represents texture metric coefficients, and β represents boundary metric coefficients;
and step 4, arranging the comprehensive differences from small to large and combining the comprehensive differences in sequence to obtain a segmentation result with a hierarchical relationship.
The invention applies the compression and merging theory in the field of image matting to image segmentation and provides a hierarchical segmentation method for cultural pattern images. Compared with a natural image, the texture of the cultural pattern image is more complex, the existing segmentation method can generate the phenomena of low speed and incomplete segmentation area during processing, and the method provides a new texture feature extraction operator to process the complex texture of the cultural pattern image, thereby improving the segmentation effect. Compared with the existing method, the method can completely divide the cultural elements in the cultural pattern image, and simultaneously displays different details of the cultural pattern through multiple layers, thereby providing technical support for the traditional cultural research.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a hierarchical segmentation method for cultural pattern images according to an embodiment of the invention.
Detailed Description
As shown in fig. 1, the hierarchical segmentation method for cultural pattern images of the present invention is characterized by comprising the following steps:
and step 1, converting the cultural pattern image into a Lab color space, and compressing pixels with similar color characteristics together, so that the image is divided into a plurality of areas. The image compression method can refer to the theory proposed by Tseng et al in Learning-based hierarchical graph for unsupervised matting and foreground estimation. In order to quickly obtain an over-color image, feature computation time is reduced, and only color features are used for affinity measurement during pixel compression. After pixel compression, the image pixels are gathered closer together, forming a plurality of small cells (regions).
Step 2, calculating likelihood ratios between adjacent areas based on color features, and combining the adjacent areas with the likelihood ratios larger than a preset threshold value to obtain an over-segmentation result, wherein a likelihood ratio calculation formula is as follows:
wherein i and j represent the sequence numbers of the regions, R i and Rj Representing the area of the neighborhood,
representation area R i Is used for the color value of the color filter,
representation area R j Mean color value of>
Mean color value representing boundary between adjacent regions, S representing the color represented by adjacent region R i And R is R j Covariance matrix of pixel color values of the composed region.
During the pixel-based puncturing and merging process, the pixels of the smooth region are quickly merged. However, for pixels around the boundary region, they will merge with pixels in the boundary direction, rather than with pixels in the neighboring smooth region. Thus, after the pixel-based puncturing and merging process, many small areas around the boundary occur, affecting the result of area compression and merging. In order to solve this problem, the present embodiment uses a residual removal method to process, and sets that the merging area around the boundary contains no more than [ N/5000], where N is the total number of image pixels, and merges the pixels in the residual area into a smooth area with the minimum adjacent color difference.
Step 3, calculating the difference of adjacent areas
3.1, calculating the color average difference between adjacent areas according to the following calculation formulas:
in the formula,Ri and Rj Representing the area of the neighborhood,
representation area R i Pixel average color value of>
Representation area R j Is used for the color values of the pixel averages.
Regions of similar color will merge together, but as regions become larger in the iterative merge process, the color variance within the regions increases and the importance of the color information decreases. Thus, in merging, the region color difference is represented by the color difference of the adjacent region boundaries. For two adjacent regions, a window of 3*3 was used to calculate the boundary average color difference:
in the formula,Bij Representing the boundary between adjacent regions, p and q representing the pixels of the regions on either side of the boundary, W p Representing a sliding window of size 5*5 for the center pixel p, B C (R i ,R j ) The number of (p, q) pairs representing the boundary,
respectively representing the brightness of the pixel p in Lab color space, the component from green to red, the component from blue to yellow,/respectively>
Respectively representing the brightness of the pixel q in Lab color space, the component from green to red, and the component from blue to yellow.
And 3.2, calculating texture differences between adjacent areas.
Compared with a natural image, the texture of the cultural pattern image is more complex, and the phenomenon of edge deletion easily occurs when the texture extraction is carried out by using a general method, so that a new texture characteristic operator is provided for calculation, and the texture characteristic operator consists of two parts: differential excitation and directional excitation.
Chroma difference D of adjacent areas AB (R i ,R j ) Obtained by calculation by the following formula:
in the formula,
respectively represent regions R i Average value of the components of the inner pixel from green to red, average value of the components from blue to yellow,>
respectively represent regions R j The average value of the components of the inner pixel from green to red, the average value of the components from blue to yellow.
Texture feature difference D of adjacent regions W (R i ,R j ) The method is calculated and obtained by the following steps:
3.2.1 pixel Point x for a Single region c Its differential excitation ζ (x c )
(x, y) represents the current pixel x c Coordinates, I, represents pixel x c The value of (a) is a preset constant, and the value of (a) is (0, 1)]；
3.2.2 pixel Point x for a Single region c Its directional excitation θ (x) c )
in the formula,xu 、x D 、x L 、x R Respectively representing pixels x c Adding the intensity values of the upper, lower, left and right pixels in eight adjacent areas around the pixel;
3.2.3 quantizing the differential and directional excitations of the region into a two-dimensional histogram of T x D, converting the two-dimensional histogram of T x D into a one-dimensional eigenvector W of TD x 1, calculating each pixel intensity using three windows of 3 x 3,5 x 5,7 x 7, respectively, obtaining three different eigenvectors W 1 ，W 2 and W3 And combining to obtain a final texture feature vector W= [ W ] 1 T ,W 2 T ,W 3 T ]；
3.2.4 calculating the neighboring region R i and Rj Texture feature differences:
in the formula,
representation area R i Texture feature vector, ">
Representation area R j Are calculated according to steps 3.2.1-3.2.3, respectively. />
The texture difference calculation formula between the adjacent regions is calculated as follows:
D T (R i ,R j )＝D AB (R i ,R j )*D W (R i ,R j )
in the formula,Ri and Rj Representing adjacent regions, D AB (R i ,R j ) Representing the chromaticity difference of adjacent regionsDifferent, D W (R i ,R j ) Representing texture feature differences for adjacent regions.
3.3, calculating the scale difference between adjacent areas, wherein the calculation formula is as follows:
wherein ,
representation area R i The number of pixels in +.>
Representation area R j The number of pixels in.
Region-scale features can speed up region merging when two small regions merge or merge one small region into a larger region. To avoid excessive dominance of large regions in the merging process, the region scale is limited using the following formula: in this step, the region scale is limited using the following formula:
and 3.4, calculating an interleaving value between adjacent areas.
Because the cultural pattern has processes of embroidery, knitting and the like in the manufacturing process, the pattern is smooth, and a plurality of color lines are interwoven together. Thus, the present document measures the degree of spatial interleaving between adjacent regions to merge fragmented regions.
In this embodiment, the interleaving value between adjacent areas is calculated according to the following formula:
in the formula,Np Representing the maximum size within a 5*5 sliding window centered on pixel pPixel color value, N q Represents the most pixel color value within a 5*5 sliding window centered on pixel q, C j Representation area R j Color value of the pixel with the largest inner part, C i Representation area R i The pixel color value with the largest inner, when a=b, δ (a, b) =1; when a+.b, δ (a, b) =0;
and 3.5, calculating the comprehensive difference between adjacent areas, wherein the calculation formula is as follows:
wherein α represents texture metric coefficients, and β represents boundary metric coefficients; in the present embodiment, α=3, β=1 is set.
And step 4, arranging the comprehensive differences from small to large and combining the comprehensive differences in sequence to obtain a segmentation result with a hierarchical relationship.
In addition to the embodiments described above, other embodiments of the invention are possible. All technical schemes formed by equivalent substitution or equivalent transformation fall within the protection scope of the invention.
Claims (7)
1. The hierarchical segmentation method for the cultural pattern image is characterized by comprising the following steps of:
step 1, converting a cultural pattern image into a Lab color space, and compressing pixels with similar color characteristics together, so that the image is divided into a plurality of areas;
step 2, calculating likelihood ratios between adjacent areas based on color features, and combining the adjacent areas with the likelihood ratios larger than a preset threshold value to obtain an over-segmentation result, wherein a likelihood ratio calculation formula is as follows:
wherein i and j represent the sequence numbers of the regions, R i and Rj Representing the area of the neighborhood,
representation area R i Mean color value of>
Representation area R j Mean color value of>
Mean color value representing boundary between adjacent regions, S representing the color represented by adjacent region R i And R is R j Covariance matrix of pixel color values of the composed region;
step 3, calculating the difference of adjacent areas
And 3.1, calculating the average difference of the colors between the adjacent areas and the average difference of the colors at the boundaries of the adjacent areas, wherein the calculation formulas are respectively as follows:
in the formula,Ri and Rj Representing the area of the neighborhood,
representation area R i Pixel average color value of>
Representation area R j Pixel average color value of B ij Representing the boundary between adjacent regions, p and q representing the pixels of the regions on either side of the boundary, W p Sliding window representing center pixel p, B C (R i ,R j ) The number of (p, q) pairs representing the boundary, < ->
Respectively representing the brightness of the pixel p in Lab color space, the component from green to red, the component from blue to yellow,/respectively>
Respectively representing the brightness of the pixel q in the Lab color space, the component from green to red, and the component from blue to yellow;
and 3.2, calculating texture differences between adjacent areas, wherein the calculation formula is as follows:
D T (R i ,R j )＝D AB (R i ,R j )*D W (R i ,R j )
in the formula,Ri and Rj Representing adjacent regions, D AB (R i ,R j ) Representing the chromaticity difference of adjacent regions, D W (R i ,R j ) Representing texture feature differences for adjacent regions;
3.3, calculating the scale difference between adjacent areas, wherein the calculation formula is as follows:
wherein ,
representation area R i The number of pixels in +.>
Representation area R j The number of pixels in;
and 3.4, calculating an interleaving value between adjacent areas, wherein the calculating formula is as follows:
in the formula,Np Representing the most pixel color value, N, within a sliding window centered on pixel p q Representing the most pixel color value, C, within a sliding window centered on pixel q j Representation area R j Color value of the pixel with the largest inner part, C i Representation area R i The pixel color value with the largest inner, when a=b, δ (a, b) =1; when a+.b, δ (a, b) =0;
and 3.5, calculating the comprehensive difference between adjacent areas, wherein the calculation formula is as follows:
wherein α represents texture metric coefficients, and β represents boundary metric coefficients;
step 4, arranging the comprehensive differences from small to large and combining the comprehensive differences in sequence to obtain a segmentation result with a hierarchical relationship;
in step 3.2, the texture feature differences D of the adjacent regions W (R i ,R j ) The method is calculated and obtained by the following steps:
3.2.1 pixel Point x for a Single region c Its differential excitation ζ (x c )
(x, y) represents the current pixel x c Coordinates, I, represents pixel x c Is a preset constant;
3.2.2 pixel Point x for a Single region c Its directional excitation θ (x) c )
in the formula,xu 、x D 、x L 、x R Respectively representing pixels x c Adding the intensity values of the upper, lower, left and right pixels in eight adjacent areas around the pixel;
3.2.3 quantizing the differential and directional excitations of the region into a two-dimensional histogram of T x D, converting the two-dimensional histogram of T x D into a one-dimensional eigenvector W of TD x 1, calculating each pixel intensity using three windows of 3 x 3,5 x 5,7 x 7, respectively, obtaining three different eigenvectors W 1 ，W 2 and W3 And combining to obtain a final texture feature vector W= [ W ] 1 T ,W 2 T ,W 3 T ]；
3.2.4 calculating the neighboring region R i and Rj Texture feature differences:
in the formula,
representation area R i Texture feature vector, ">
Representation area R j Are calculated according to steps 3.2.1-3.2.3, respectively.
2. The hierarchical segmentation method for cultural pattern images as defined in claim 1, wherein: in step 2, processing the over-segmentation result by adopting a residual error removing method, setting the merging area around the boundary to contain no more than [ N/5000], wherein N is the total number of image pixels, and merging pixels in the residual error area into a smooth area with minimum adjacent chromatic aberration.
3. The hierarchical segmentation method for cultural pattern images as defined in claim 1, wherein: in step 3.3, the region scale is limited using the following formula:
4. the hierarchical segmentation method for cultural pattern images as defined in claim 1, wherein: in step 3.2, the chrominance difference D of the adjacent areas AB (R i ,R j ) Obtained by calculation by the following formula:
in the formula,
respectively represent regions R i Average value of the components of the inner pixel from green to red, average value of the components from blue to yellow,>
respectively represent regions R j The average value of the components of the inner pixel from green to red, the average value of the components from blue to yellow.
5. The hierarchical segmentation method for cultural pattern images as defined in claim 1, wherein: in step 3.2.1, the value of σ is in the range of (0, 1).
6. The hierarchical segmentation method for cultural pattern images as defined in claim 1, wherein: in step 3.5, α=3, β=1 is set.
7. The hierarchical segmentation method for cultural pattern images as defined in claim 1, wherein: in step 3, the size of the sliding window is 5*5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110563186.1A CN113160252B (en) | 2021-05-24 | 2021-05-24 | Hierarchical segmentation method for cultural pattern image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110563186.1A CN113160252B (en) | 2021-05-24 | 2021-05-24 | Hierarchical segmentation method for cultural pattern image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113160252A CN113160252A (en) | 2021-07-23 |
| CN113160252B true CN113160252B (en) | 2023-04-21 |
Family
ID=76877535
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110563186.1A Active CN113160252B (en) | 2021-05-24 | 2021-05-24 | Hierarchical segmentation method for cultural pattern image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113160252B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107833224A (en) * | 2017-10-09 | 2018-03-23 | 西南交通大学 | A kind of image partition method based on multi-level region synthesis |
| CN109272467A (en) * | 2018-09-25 | 2019-01-25 | 南京大学 | A kind of stratification image partition method based on multi-scale edge clue |
-
2021
- 2021-05-24 CN CN202110563186.1A patent/CN113160252B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107833224A (en) * | 2017-10-09 | 2018-03-23 | 西南交通大学 | A kind of image partition method based on multi-level region synthesis |
| CN109272467A (en) * | 2018-09-25 | 2019-01-25 | 南京大学 | A kind of stratification image partition method based on multi-scale edge clue |
Non-Patent Citations (3)
| Title |
|---|
| Fractal Beauty in Xinjiang Folk Art Patterns;Peng Hong等;《Computer Aided Drafting, Design and Manufacturing》;20140930;全文 * |
| 基于边缘先验的文化图案层次分割算法;寇晓斌 等;《电子技术与软件工程》;20210901;全文 * |
| 基于边缘形态变换的彩色织物图像分割算法;赵海英等;《中国体视学与图像分析》;20111231(第01期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113160252A (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111415363B (en) | Image edge identification method | |
| Luan et al. | Natural image colorization | |
| US8774503B2 (en) | Method for color feature extraction | |
| CN108537239B (en) | Method for detecting image saliency target | |
| CN102800094A (en) | Fast color image segmentation method | |
| CN110458172A (en) | A kind of Weakly supervised image, semantic dividing method based on region contrast detection | |
| CN112101370B (en) | Automatic image matting method for pure-color background image, computer-readable storage medium and equipment | |
| CN111583279A (en) | Super-pixel image segmentation method based on PCBA | |
| CN110268442B (en) | Computer-implemented method of detecting a foreign object on a background object in an image, device for detecting a foreign object on a background object in an image, and computer program product | |
| CN111932601B (en) | Dense depth reconstruction method based on YCbCr color space light field data | |
| CN108038458B (en) | Method for automatically acquiring outdoor scene text in video based on characteristic abstract diagram | |
| CN114511567B (en) | Tongue body and tongue coating image identification and separation method | |
| CN114708165A (en) | Edge perception texture filtering method combining super pixels | |
| CN113989299A (en) | Open-pit mine rock stratum image segmentation method based on k-means clustering | |
| CN105608683B (en) | A kind of single image to the fog method | |
| Tan et al. | Image haze removal based on superpixels and Markov random field | |
| CN115100226A (en) | Contour extraction method based on monocular digital image | |
| CN111868783B (en) | Region merging image segmentation algorithm based on boundary extraction | |
| CN113160252B (en) | Hierarchical segmentation method for cultural pattern image | |
| JP3923243B2 (en) | Character extraction method from color document image | |
| Shan et al. | Image highlight removal based on double edge-preserving filter | |
| CN112560740A (en) | PCA-Kmeans-based visible light remote sensing image change detection method | |
| CN111915500A (en) | Foggy day image enhancement method based on improved Retinex algorithm | |
| CN115601358B (en) | Tongue picture image segmentation method under natural light environment | |
| CN108154485A (en) | A kind of ancient painting restorative procedure based on layering and stroke direction parsing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |